Sound Source Separation Filters for Low-Delay Hearing Aid Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional sound source separation techniques face challenges in reducing delay time, making them unsuitable for applications like hearing aids where a time lag of over 10 ms is unacceptable, especially when performing online processing.
Innovation Solution
The sound source separation apparatus employs a separation-matrix processor that updates the separation matrix in the frequency domain and partially removes non-causal components from time-series filter coefficients, allowing for real-time processing with significantly reduced delay time by using a filter-coefficient transformer and FIR filters, ensuring efficient sound separation without deteriorating performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional sound source separation techniques are used for online processing, then sound separation can be performed in real-time, but the delay time exceeds 10 ms which is unacceptable for hearing aids
Solution Approach 1:
The patent segments the filter coefficients into causal and non-causal components. By separating these components and only applying the causal part in online processing, the system achieves real-time operation with reduced delay time while maintaining effective sound separation functionality.
Solution Approach 2:
The patent extracts and removes the non-causal components from the filter coefficients before applying them in the separator. This extraction allows the system to eliminate the time delay introduced by non-causal filtering while preserving the essential separation performance through the retained causal components.
2Loss of time
If non-causal components are removed from filter coefficients, then delay time is reduced, but sound separation performance may deteriorate
Solution Approach 1:
The patent applies partial action by using only the causal portion of the filter coefficients rather than the complete filter including non-causal components. This partial application is sufficient to achieve acceptable sound separation performance while dramatically reducing the delay time to within the 10 ms threshold for hearing aid applications.
Data Source
AI summary
A sound source separation apparatus includes: a separation-matrix processor that transforms a plurality of observation signals corresponding to sounds being propagated from a plurality of sound sources into a frequency-domain signal group the separation-matrix processor updating a separation matrix based on the frequency-domain signal group and transforming the updated separation matrix into time-series filter coefficients to output; a filter-coefficient transformer that partially removes non-causal components from the filter coefficients to transform the filter coefficients, and a separator that supplies the filter coefficients to a filter group, the separator generating a plurality of separation signals separated from the plurality of observation signals corresponding to the separation matrix.